PROJECT SUMMARY The workforce of physician scientists, particularly in positions of senior leadership, persistently fails to reflect the diversity seen in both the general population and the diversity in US medical schools, with women and minorities underrepresented in the physician-scientist pipeline. One potential mechanism to address this is the presence of scholarly concentration programs (SCPs) within the traditional 4-year medical school curriculum. SCPs allow MD-only degree candidates to participate in a mentored research projects in an effort to increase interest in career-long research. SCP outcomes are often measured by increased publication rates or increased intention for a research career from matriculation to graduation, which has been shown to predict successful entry into a physician-scientist career. The Scholarly Concentrations Collaborative is a coalition of 10 medical schools actively conducting research on mentoring and promotion of scientific careers. Notably, the Collaborative has undertaken multi-site research to study the predictors of increased intention to enter research careers related to mentorship. While SCPs have primarily focused on program evaluation to date, theoretical and mechanistic studies exploring the science of mentoring in SCPs are lacking. Social cognitive career theory (SCCT) is one theoretical model that could explain successful mentoring of medical students into future physician scientists. Several studies indicate that research self-efficacy is an important predictors of entry into scientific careers. At University of Chicago, we found that research self-efficacy at matriculation is associated with increased intent for career-long research when controlling for initial interest in basic science, gender, and minority status. In a 2-site study of first year students at the University of Chicago and the Icahn School of Medicine at Mount Sinai, we found that establishment of a strong mentoring relationship is rated as the most important goal among medical students, and is valued more by female students compared to male peers. Lastly, in a multi-site study of mentoring among graduates from 10 institutions, we found that although women and URM students are more likely to select mentors who are gender- or minority-status concordant, mentor-student concordance is not associated with intention to enter a research career or publication. Instead, we found that mentors who invested in the professional development of their students (compared to those that invested in project completion) were rated most impactful on students' intent for career-long research and students' professional identity. Based on our preliminary data, we will conduct a randomized trial of 300 mentors of women and minority medical student mentees to test whether a targeted intervention based on social cognitive career theory for mentors could lead to increased research self-efficacy, research career persistence, and increased objective measures of research productivity among their mentees. In addition to understanding whether such an intervention could succeed, it is equally important to understand the underlying mechanisms for how and why a mentoring intervention based on SCCT succeeds. We will specifically examine whether trainee engagement, as measured by Experience Sampling Method in real-time, mediates the effects of the intervention on relevant research career outcomes.